A finitary Kronecker's lemma and large deviations in the Strong Law of Large numbers on Banach spaces

Morenikeji Neri

2024-11-13

Abstract:We explore the computational content of Kronecker's lemma via the proof-theoretic perspective of proof mining and utilise the resulting finitary variant of this fundamental result to provide new rates for the Strong Law of Large Numbers for random variables taking values in type $p$ Banach spaces, which in particular are very uniform in the sense that they do not depend on the distribution of the random variables. Furthermore, we provide computability-theoretic arguments to demonstrate the ineffectiveness of Kronecker's lemma and investigate the result from the perspective of Reverse Mathematics. In addition, we demonstrate how this ineffectiveness from Kronecker's lemma trickles down to the Strong Law of Large Numbers by providing a construction that shows that computable rates of convergence are not always possible. Lastly, we demonstrate how Kronecker's lemma falls under a class of deterministic formulas whose solution to their Dialectica interpretation satisfies a continuity property and how, for such formulas, one obtains an upgrade principle that allows one to lift computational interpretations of deterministic results to quantitative results for their probabilistic analogue. This result generalises the previous work of the author and Pischke.

Logic,Probability

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to explore the computational content of Kronecker's lemma from the perspective of proof mining, and use the resulting finite variants to provide new convergence rates for the strong law of large numbers of random variables in Banach spaces. Specifically:

1. **Computational content of Kronecker's lemma**:
- Through the method of proof mining, the paper studies the computational content of Kronecker's lemma. The author proves a quantitative finite version of Kronecker's lemma and uses its Dialectica interpretation to solve this problem.
- The author also shows the computational ineffectiveness of Kronecker's lemma, that is, in some cases, a computational convergence rate cannot be obtained.

2. **Convergence rates of the strong law of large numbers**:
- Using the above - mentioned finite version of Kronecker's lemma, the author provides new convergence rates for the strong law of large numbers of random variables in Banach spaces. These convergence rates are very uniform and do not depend on the distribution of random variables.
- The author further explores the significance of these convergence rates in reverse mathematics and proves that the existence rate of Kronecker's lemma is equivalent to ACA_0.

3. **Applications in probability theory**:
- The paper also studies the probability - theory version of Kronecker's lemma and its application in the strong law of large numbers. The author provides a computational interpretation, showing how to extract quantitative results from deterministic results and generalize them to probability - theory results.
- Finally, by constructing examples, the author shows that in some cases, a computational almost - sure convergence rate does not exist, thus further proving the limitation of the effectiveness of Kronecker's lemma.

In summary, the main objective of this paper is to deeply study the computational content of Kronecker's lemma through the method of proof mining, and apply it to the strong law of large numbers of random variables in Banach spaces, providing new convergence rates and computational interpretations.